The Central Functions of Cities in Poland in Light of Administrative Reform

Abstract

The aim of this article is to determine and assess the changes in the central functions of major Polish cities in the context of Poland’s 1999 administrative reform. This study covers a period of twenty-two years (1998–2019), which provides a solid perspective for evaluating the functional changes occurring within the current territorial division of the country. The conclusions indicate that the dynamics of changes in central functions in former provincial capitals differed from those in cities that retained this status. At the same time, it was found that the administrative reform sanctioned the scope and scale of earlier socio-economic dependencies in the present regional centers. The research results suggest that administrative reforms in Poland solidified existing socio-economic development paths in cities, but were not the sole catalyst for change. They point to a complex interaction of factors influencing urban evolution. By utilizing indicators that measure the centrality of cities in relation to their demographic potential, this research offers a quantitative assessment of the level of development of central functions, which is crucial for monitoring progress towards sustainable urbanization. By providing this information, this research also contributes to a broader discussion on the advancement and dynamics of sustainable urban development.

1. Introduction

The socio-economic trajectories of cities are profoundly shaped by both structural reforms and the inherent characteristics of urban systems. This study examines the impact of the 1999 administrative reform in Poland, which redefined the status of regional capitals and altered the central functions of many cities. Focusing on the evolution of urban hierarchies and economic roles, this study situates its findings within the broader context of economic and spatial theories.

Despite the significant number of studies on this topic, there remains a clear research gap in understanding the long-term effects of losing administrative status on the dynamics of urban development. Many studies focus on short-term effects or residents’ subjective perceptions, neglecting more detailed analyses of economic and demographic data. Furthermore, much research concentrates on cities that have lost administrative functions without simultaneously conducting a comparative analysis of cities that have retained their status.

The example of Poland, where the 1999 administrative reform reduced the number of voivodeships (administrative regions) from 49 to 16 [1] and changed the status of many cities, provides a unique case study for analyzing the impact of such reforms on urban development.

This article aims to fill this gap by analyzing how the loss of administrative function affected urban development in Poland, both in the short and long term. By using centrality indicators and other socio-economic measures, this study assesses whether losing the status of regional capital leads to a significant slowdown in urban development and identifies factors that may mitigate or exacerbate these effects. The results of this analysis provide valuable insights for policymakers, urban planners, and researchers seeking to understand and manage the dynamic processes of urban transformation.

The aim of the article is to determine and evaluate changes in the implementation of central functions in major Polish cities in the context of the 1999 administrative reform. This study covers a twenty-two-year period (1998–2019), providing a good perspective for assessing functional changes within the current territorial division. The research encompasses the 49 cities (Figure 1) that served as provincial capitals until 1998. These can be divided into two groups.

The first group includes the cities that retained this status after 1998: Białystok, Bydgoszcz, Gdańsk, Gorzów Wielkopolski, Katowice, Kielce, Kraków, Lublin, Łódź, Olsztyn, Opole, Poznań, Rzeszów, Szczecin, Toruń, Warsaw, Wrocław, and Zielona Góra. Fourteen of these cities serve as both the voivode’s seat and the regional self-government’s seat. In two voivodeships, these administrative functions are shared: in the Kuyavian-Pomeranian Voivodeship, Bydgoszcz is the voivode’s seat, while Toruń houses the regional assembly and marshal’s office; similarly, in the Lubusz Voivodeship, Gorzów Wielkopolski is the voivode’s seat, and Zielona Góra is the self-government seat. As a result, 18 cities play the role of a regional capital in the 16 Polish voivodeships.

The second group consists of the 31 cities that lost their status as a provincial capital due to the reform: Biała Podlaska, Bielsko-Biała, Chełm, Ciechanów, Częstochowa, Elbląg, Jelenia Góra, Kalisz, Konin, Koszalin, Krosno, Legnica, Leszno, Łomża, Nowy Sącz, Ostrołęka, Piła, Piotrków Trybunalski, Płock, Przemyśl, Radom, Siedlce, Sieradz, Skierniewice, Słupsk, Suwałki, Tarnobrzeg, Tarnów, Wałbrzych, Włocławek, Zamość. Except for Ciechanów, Piła, Sieradz, and Wałbrzych (temporarily), these cities retained the status of being a powiat (metropolitan district) [2].

This study will investigate the hypothesis that administrative changes influence the process of urban development, in particular that losing provincial capital status limits this development.

2. Theoretical Frameworks

Urban theory has long grappled with the forces shaping city development and regional dynamics. Walter Christaller’s central place theory, a milestone in spatial economics, posits that cities function as central hubs providing goods and services to surrounding areas, thereby creating a hierarchical urban network [3]. This theory suggests that the centrality of a place, defined by the concentration of a range of services, plays a crucial role in its economic vitality and socio-political significance.

Christaller’s model faced criticism for its simplified view of urban systems, especially its static description of the distribution of population and the provision of services. However, later researchers, including Brian Berry and August Lösch, expanded on these ideas, integrating them with empirical observations to give a better understanding of patterns of urban and regional development [4,5].

In the second half of the 20th century, spatial economists and urbanists began to focus more on the dynamic aspects of urban growth, introducing concepts such as agglomeration economies and urban scaling laws. Some assumptions of Christaller were particularly questioned due to their static nature [6] and simplified understanding of the hierarchy of cities based solely on population size [7,8,9].

Paul Krugman and Masahisa Fujita further developed these ideas within the framework of new economic geography. Although this concept arises from somewhat different theoretical constructs, such as François Perroux’s theory of growth poles [10], based on external economies of scale and agglomeration, or Paul Samuelson’s theory of location and international trade [11], their work emphasized the role of increasing returns to scale and transport costs in shaping the spatial distribution of economic activity [12]. They introduced models explaining how agglomerative forces—such as economies of scale, network effects, and labor market pooling—compete with dispersive forces, such as congestion and the high cost of land, to determine the size and spatial arrangement of cities. A similar concept in their model of the new economic geography of central places, combining Christaller’s ideas with those of Fujita and Mori, was presented in [13].

Other scholars [14,15] have also contributed to the development and adaptation of classical central place theories, refining them to suit contemporary urban realities. Parr, in his analysis of multi-tiered urban structures, highlights the need to integrate the theory with a more flexible approach that considers polycentricity and the functional diversity of cities. On the other hand, [15] expands Christaller’s theories within the context of the growing importance of global cities and international economic networks, leading to a reassessment of traditional assumptions regarding spatial hierarchies.

In the context of contemporary research, it is also important to highlight the role of new analytical tools such as GIS systems and AI-based models, which allow for more precise analysis of spatial structures and enable a better understanding of the complex interactions between administrative, economic, and social factors. For example, the works [16,17] integrate traditional theories with modern approaches based on big data analysis and simulation models. As a result, central place theory, despite its limitations, finds new applications, particularly in studies on dynamic processes of regional transformation and sustainable development.

In this way, W. Christaller’s central place theory, understood as a spatial-economic model, was revitalized. Its classic form still holds significance in the social sciences and spatial economics. Its simplicity, while not fully capturing the complexity of modern urban systems, nevertheless provides a universal analytical framework. In the context of the increasing availability of spatial data and GIS analysis tools, this theory is finding new applications and can be used as an element of more advanced research models. This has resulted in a certain renaissance of central place theory. Fujita, Krugman, and Mori [18] even stated that this theory was simply too powerful to remain undeveloped.

The utility of this concept for studying post-socialist realities was demonstrated by [19]. It examined how the political and economic changes that followed the fall of communism impacted the hierarchy of central places in Poland and other countries in the region. These analyses included factors such as administrative decentralization, changes in economic structure, and population migration. It has been argued that Christaller’s theory could be useful in analyzing these processes, provided it is adapted to specific historical and social conditions.

Similarly, the work [20] refers to central place theory, adapting it to contemporary socio-economic realities. Like [19], it emphasizes the need to consider specific local and regional conditions, as well as new trends such as sustainable development and polycentrism, in the analysis of urban structures. Its contributions highlight the ongoing evolution of Christaller’s theory and its adaptation to changing global conditions.

In the context of research on the impact of administrative factors on urban development trajectories, central place theory offers valuable analytical tools for understanding how changes in administrative structures can affect the hierarchy and functions of cities. Administrative reforms, such as changes in the status of regional capitals or the redistribution of government functions, have the potential to transform local economies, influencing their ability to attract investment, population, and capital.

The literature often highlights that granting a city administrative functions can lead to multiplier effects that promote development in other areas of the urban economy. This may form a basis for the future development of a city [21,22,23,24,25]. However, some authors also note negative aspects of being a regional or national capital. For example, the dominance of public functions may lead to a drain of talent and resources, which negatively impacts development in other areas of activity [25,26,27]. Another threat might lie in the rise of a bureaucratic culture of administration and management, leading to an inflexible approach that acts as a barrier to entrepreneurship and innovation [28].

A comprehensive review of empirical studies based on the theories described above can be found in [29,30,31]. It is worth discussing several fundamental works in more detail, especially as contemporary research on the impact of the loss of administrative functions on urban development trajectories emphasizes the diverse effects of this process.

For example, research conducted in Austria [32] showed that the loss of the status of being a regional capital leads to a significant decline in public and private investment, often resulting in long-term economic regression. Similar results appear in analyses concerning cities in Germany [33], where it was observed that cities losing administrative functions experience greater outmigration and loss of talent, exacerbating the marginalization of these centers. These studies, continuing the fundamental work [23], showed that loss of administrative status could lead to increasing unemployment, decreasing incomes, and the weakening of the service sector, which is often closely linked to the presence of administrative institutions.

However, research in Poland [34] indicates that the perception of marginalization may not always correspond to economic reality. These analyses show that while cities losing the status of regional capital often report concerns about declining prestige and significance, actual socio-economic indicators do not always show dramatic regression. In some cases, these cities, despite losing administrative functions, find alternative development paths, such as specializing in innovative or tourism sectors.

Additionally, a study [35], based on analyses conducted in several European countries, suggests that the effects of losing administrative functions may be less straightforward and rather dependent on the regional context and structural conditions of particular cities. It indicates that a key factor determining the consequences of losing administrative status is the prior level of economic diversification and the degree of involvement of local government in restructuring processes and seeking new sources of economic growth.

Similar studies outside Europe also reveal the complexity of these processes. For example, research in Australia [36] showed that cities losing the status of state capital face challenges related to halting the outflow of inhabitants and decline in investment, often leading to regional inequalities. Similarly, studies in the United States [37] indicate that changes in administrative status can affect public policy and the redistribution of resources, which in turn impacts a city’s ability to adapt and pursue sustainable development.

These observations suggest that administrative reform can act as a catalyst for change, influencing urban development trajectories by modifying the central functions of cities. In practice, this means that cities often need to adapt their development strategies to respond to changing administrative and economic conditions. In summary, research on the impact of administrative factors on urban development uses central place theory as an analytical framework, enabling an understanding of how political and administrative decisions shape urban economic and social landscapes. Integrating these theories with modern approaches and empirical research enables a better understanding of the complex interactions that define the development of contemporary cities.

3. Materials and Methods

The level of development of the central functions of a city was measured using two indicators: the indicator of the centrality of a city ${(W}_C)$ and the indicator of the centrality of a city corrected according to its demographic potential ${(W}_C^M)$. It should be noted here that a similar approach was used by Przybyła and Kachniarz [29]. The data used there covered the period from 1998 to 2013. The study presented here uses those data, together with new data covering the period up to 2019.

The indicator of the centrality of a city, W_C (Equation (1)), measures the number of people employed in a given sector within a city relative to the number of people employed in that sector in the administrative region containing that city. This term is the difference between the number of people employed in a sector within a city and the number of people that would be employed if the proportion employed in that sector was the same as in the administrative region. When the proportion of individuals employed in a sector within a city is fixed, this indicator is proportional to the population of the city. Hence, in order to compare cities with different populations, it is necessary to scale W_C according to the population of a city. The scaled indicator of the centrality of a city corrected according to its demographic potential, $W_C^M$ (Equation (2)), measures the relative excess of the number employed in a sector within a city per 1000 inhabitants. A large positive value of this indicator shows that a city exhibits specialization in that sector (high centrality), while low or negative values show a low level of specialization (low centrality). A large value of the indicator for general employment shows that a relatively large number of people commute into the city. Hence, this indicator enables us to compare the centrality of cities and their influence on the surrounding region [38].

• Equation (1) shows the indicator of the centrality of a city:

$$W_C=Z_{iM}-L_M{\displaystyle \frac{Z_{i0}}{L_0}}$$

(1)

where the following definitions hold:

$W_C$—indicator of the centrality of a city;

$Z_{iM}$—number employed in Sector i within a city;

$L_M$—population of a city;

$Z_{i0}$—number employed in Sector i within the region of attraction (administrative region) of a city;

$L_0$—population of the region of attraction (administrative region) of a city.

• Equation (2) shows the indicator of the centrality of a city corrected according to its population:

$$W_C^M={\displaystyle \frac{W_C}{L_M}}·1000$$

(2)

where the following definitions hold:

$W_C^M$—indicator of the centrality of a city corrected according to its population;

$W_C$—indicator of the centrality of a city (based on Equation (1));

$L_M$—population of a city.

It should be noted that, as in previously conducted studies, the region of attraction of each city is assumed to correspond to the set of local government administrations included in the original region (voivodeship). It should be noted that under the administrative reforms of 1998, in general, several regions were combined to form larger regions. For example, the Jelenia Góra, Legnica, Wałbrzych, and Wrocław regions were combined into the new region of Lower Silesia with its capital in Wrocław. The changes in the borders of administrative regions at lower levels were minimal. This enables a comparison of the central functions of a city at different points in time; see [29].

It should also be noted that the classification of businesses by the Polish government changed between 1998 and 2019. The Polish Business Classification of 2007 (PBC 2007) has been used since 2008, replacing the PBC 2004 system. These changes mean that some of the classes are not comparable. Hence, it was necessary to use expert knowledge to correct for the differences between these data; see [29].

In order to analyze the relationship between the relative measure of centrality, $W_C^M,$ or changes in the global value of a city’s indicator of centrality (adjusted for population size) and the population of a city (measured in 1998, the start of the study period), Spearman’s coefficient of correlation, r_S, was used. This coefficient takes values between −1 and +1 (inclusively). Values close to zero indicate a lack of a systematic relation between the population of a city and its indicator of centrality. When r_S = 1, the indicator of centrality is strictly increasing in a city’s population. When r_S = −1, the indicator of centrality is strictly decreasing in a city’s population. In order to assess the strength of association between two variables, we use the following scale:

$\left|r_S\right|<0.3$—weak (or no) dependence;

$0.3\le \left|r_S\right|<0.6$—moderate dependence;

$\left|r_S\right|\ge 0.6$—strong dependence.

It should be noted that the adjusted index of centrality, $W_C^M$, does not have a normal distribution and the number of cities investigated is not very large. For this reason, in order to assess the relations between the status of a city (current or former capital) and the adjusted index of centrality (including changes in the index of centrality), a non-parametric test was used, the Wilcoxon rank test.

4. Changes in the Central Function of Polish Cities

Studies on the centrality of cities have concentrated on the various sectors of service provided in the cities considered; see [18,23,38,39,40]. Such data are presented in

Table 1. The number employed in service sectors in 1998 according to PBC 2004, together with population sizes.

Territorial Unit	Total	Including			Population
		Service Sector—Total	Service Sector—Private Enterprises *	Public Service Sector **
ŁÓDZKIE	608,996	326,344	168,402	157,942	2,663,608
Łódź	240,293	145,084	82,234	62,850	806,728
Piotrków Trybunalski	28,342	15,375	8079	7296	81,284
Sieradz	16,117	10,993	5106	5887	45,529
Skierniewice	14,938	9212	4881	4331	48,740
MAZOWIECKIE	1,406,035	904,939	581,383	323,556	5,066,598
Ciechanów	16,050	10,132	4904	5228	47,468
Płock	51,555	23,750	13,720	10,030	131,011
Ostrołęka	19,953	11,283	5919	5364	55,271
Siedlce	26,853	17,198	8595	8603	75,613
Radom	62,851	36,804	19,818	16,986	232,262
Warsaw	781,623	564,900	412,596	152,304	1,618,468
MAŁOPOLSKIE	722,142	416,452	222,371	194,081	3,215,885
Cracow	288,541	180,232	107,404	72,828	740,666
Nowy Sącz	34,170	21,285	12,456	8829	83,754
Tarnów	51,959	28,276	16,081	12,195	121,494
ŚLĄSKIE	1,410,112	673,002	405,953	267,049	4,882,448
Bielsko-Biała	72,921	36,824	23,505	13,319	180,307
Częstochowa	83,843	46,182	26,471	19,711	257,812
Katowice	184,127	109,807	75,665	34,142	345,934
LUBELSKIE	419,579	259,086	121,191	137,895	2,239,500
Biała Podlaska	18,549	12,282	6061	6221	58,351
Chełm	20,799	13,610	7222	6388	70,654
Zamość	22,649	15,526	8046	7480	68,288
Lublin	123,244	82,256	42,531	39,725	356,251
PODKARPACKIE	444,778	237,301	114,603	122,698	2,122,164
Krosno	28,997	13,407	6551	6856	49,227
Przemyśl	23,489	15,935	7618	8317	68,455
Rzeszów	74,437	45,439	26,130	19,309	162,049
Tarnobrzeg	15,355	9127	4174	4953	51,061
PODLASKIE	229,524	143,559	69,241	74,318	1,223,802
Białystok	88,579	59,108	32,720	26,388	283,937
Łomża	16,296	11,048	5604	5444	64,605
Suwałki	20,084	12,075	5729	6346	68,331
ŚWIĘTOKRZYSKIE	265,151	145,176	70,490	74,686	1,326,218
Kielce	79,249	48,584	28,116	20,468	212,383
LUBUSKIE	240,072	134,977	72,183	62,794	1,022,521
Gorzów Wielkopolski	46,689	25,075	14,477	10,598	126,019
Zielona Góra	42,990	29,284	17,995	11,289	118,182
WIELKOPOLSKIE	848,347	439,320	253,555	185,765	3,351,422
Kalisz	41,271	21,155	12,315	8840	106,641
Konin	32,356	17,166	10,001	7165	83,426
Leszno	22,122	13,895	8491	5404	62,274
Piła	27,876	16,696	10,046	6650	76,849
Poznań	237,258	154,651	99,601	55,050	578,235
ZACHODNIOPOMORSKIE	414,458	242,680	135,063	107,617	1,731,804
Koszalin	35,975	24,096	13,440	10,656	112,375
Szczecin	143,844	92,049	58,763	33,286	416,988
DOLNOŚLĄSKIE	730,234	405,529	213,505	192,024	2,982,128
Jelenia Góra	29,117	19,228	10,183	9045	93,901
Legnica	33,945	21,853	12,649	9204	109,335
Wałbrzych	34,246	20,934	11,260	9674	136,923
Wrocław	215,092	141,588	83,061	58,527	637,877
OPOLSKIE	248,501	123,339	63,600	59,739	1,089,586
Opole	53,524	34,369	20,064	14,305	129,553
KUJAWSKO-POMORSKIE	483,687	259,696	135,942	123,754	2,100,106
Bydgoszcz	128,521	72,937	44,055	28,882	386,855
Toruń	75,096	41,703	24,545	17,158	206,158
Włocławek	39,971	19,823	10,769	9054	123,373
POMORSKIE	521,713	304,478	173,462	131,016	2,185,673
Słupsk	31,460	19,604	10,740	8864	102,370
Gdańsk	151,789	99,386	59,577	39,809	458,988
WARMIŃSKO-MAZURSKIE	307,338	177,617	90,386	87,231	1,463,520
Elbląg	37,636	21,660	11,561	10,099	129,782
Olsztyn	64,278	43,305	25,931	17,374	170,904

* this sector comprises the following: wholesale and retail trade, servicing of private vehicles and household appliances, hotels and restaurants, transport, storage, financial services, real estate services, communal, social and individual services, employment in households, international organizations. ** this sector comprises the following: public administration and national defense, obligatory social services, education, health and social care. Source: [29].

,

Table 2. The centrality measures of Polish cities in 1998 (corrected for population size).

Territorial Unit	Global Measure of Centrality (Corrected for Population Size)	Component Measures of Centrality
		Service Sector—Private Enterprises	Public Service Sector
ŁÓDZKIE
Łódź	57.3	38.7	18.6
Piotrków Trybunalski	66.6	36.2	30.5
Sieradz	118.9	48.9	70.0
Skierniewice	66.5	36.9	29.6
MAZOWIECKIE
Ciechanów	34.8	−11.4	46.3
Płock	2.7	−10.0	12.7
Ostrołęka	25.5	−7.7	33.2
Siedlce	48.8	−1.1	49.9
Radom	−20.1	−29.4	9.3
Warsaw	170.4	140.2	30.2
MAŁOPOLSKIE
Cracow	113.8	75.9	38.0
Nowy Sącz	124.6	79.6	45.1
Tarnów	103.2	63.2	40.0
ŚLĄSKIE
Bielsko-Biała	66.4	47.2	19.2
Częstochowa	41.3	19.5	21.8
Katowice	179.6	135.6	44.0
LUBELSKIE
Biała Podlaska	94.8	49.8	45.0
Chełm	76.9	48.1	28.8
Zamość	111.7	63.7	48.0
Lublin	115.2	65.3	49.9
PODKARPACKIE
Krosno	160.5	79.1	81.5
Przemyśl	121.0	57.3	63.7
Rzeszów	168.6	107.2	61.3
Tarnobrzeg	66.9	27.7	39.2
PODLASKIE
Białystok	90.9	58.7	32.2
Łomża	53.7	30.2	23.5
Suwałki	59.4	27.3	32.1
ŚWIĘTOKRZYSKIE
Kielce	119.3	79.2	40.1
LUBUSKIE
Gorzów Wielkopolski	67.0	44.3	22.7
Zielona Góra	115.8	81.7	34.1
WIELKOPOLSKIE
Kalisz	67.3	39.8	27.5
Konin	74.7	44.2	30.5
Leszno	92.0	60.7	31.3
Piła	86.2	55.1	31.1
Poznań	136.4	96.6	39.8
ZACHODNIOPOMORSKIE
Koszalin	74.3	41.6	32.7
Szczecin	80.6	62.9	17.7
DOLNOŚLĄSKIE
Jelenia Góra	68.8	36.8	31.9
Legnica	63.9	44.1	19.8
Wałbrzych	16.9	10.6	6.3
Wrocław	86.0	58.6	27.4
OPOLSKIE
Opole	152.1	96.5	55.6
KUJAWSKO-POMORSKIE
Bydgoszcz	64.9	49.1	15.7
Toruń	78.6	54.3	24.3
Włocławek	37.0	22.6	14.5
POMORSKIE
Słupsk	52.2	25.6	26.6
Gdańsk	77.2	50.4	26.8
WARMIŃSKO-MAZURSKIE
Elbląg	45.5	27.3	18.2
Olsztyn	132.0	90.0	42.1

Source: [27].

,

Table 3. The number employed in various service sectors in 2013 according to PBC 2007, together with population sizes.

Territorial Unit	Total	Including			Population
		I *	II **	III ***
ŁÓDZKIE	738,806	125,879	24,606	205,534	2,513,093
Łódź	230,821	56,905	14,861	97,896	711,332
Piotrków Trybunalski	24,332	5289	864	9377	75,903
Sieradz ****	12,323	2391	429	5458	43,356
Skierniewice	12,300	1947	451	4898	48,634
MAZOWIECKIE	1,759,966	429,425	129,724	572,966	5,316,840
Ciechanów ****	13,246	2869	535	5158	44,797
Płock	46,386	10,112	1646	16,074	122,815
Ostrołęka	16,850	3290	612	6326	52,917
Siedlce	24,998	5917	610	9838	76,347
Radom	54,009	12,538	1526	21,364	218,466
Warsaw	828,319	259,399	109,285	349,581	1,724,404
MAŁOPOLSKIE	997,938	186,646	29,887	282,147	3,360,581
Cracow	306,706	86,407	19,859	133,185	758,992
Nowy Sącz	32,421	9348	972	10,550	83,943
Tarnów	41,151	9310	1232	15,390	112,120
ŚLĄSKIE	1,286,080	253,298	49,142	380,160	4,599,447
Bielsko-Biała	72,590	16,202	1951	22,760	173,699
Częstochowa	73,860	16,382	2482	25,401	232,318
Katowice	164,089	42,014	16,055	68,417	304,362
LUBELSKIE	677,862	82,397	16,361	161,655	2,156,150
Biała Podlaska	15,239	4155	583	5883	57,658
Chełm	15,447	3496	645	6353	65,481
Zamość	19,597	5103	990	8399	65,255
Lublin	122,708	29,876	8584	56,526	343,598
PODKARPACKIE	681,212	87,424	11,921	158,932	2,129,294
Krosno	24,845	5652	533	7962	47,223
Przemyśl	18,631	4150	696	8516	63,638
Rzeszów	85,674	20,660	4254	34,699	183,108
Tarnobrzeg	12,626	2078	467	5280	48,217
PODLASKIE	335,804	49,448	8355	87,988	1,194,965
Białystok	85,952	22,870	4183	38,163	295,282
Łomża	15,701	3826	641	6958	62,711
Suwałki	19,530	4111	984	6245	69,317
ŚWIĘTOKRZYSKIE	377,567	46,975	6838	94,031	1,268,239
Kielce	75,195	18,369	3120	32,271	199,870
LUBUSKIE	248,282	46,921	7211	75,797	1,021,470
Gorzów Wielkopolski	40,759	8012	1888	15,466	124,344
Zielona Góra	40,767	11,696	2268	17,299	118,405
WIELKOPOLSKIE	1,061,886	209,955	32,254	264,605	3,467,016
Kalisz	34,043	8095	1030	10,567	103,997
Konin	24,639	6140	935	8716	77,224
Leszno	21,492	5472	934	6749	64,589
Piła ****	22,774	5905	1100	7759	74,609
Poznań	231,362	64,784	16,294	98,934	548,028
ZACHODNIOPOMORSKIE	370,141	79,891	13,173	126,473	1,718,861
Koszalin	31,860	8642	1290	11,987	109,170
Szczecin	110,622	33,039	6597	48,017	408,172
DOLNOŚLĄSKIE	794,138	155,723	33,132	259,581	2,909,997
Jelenia Góra	24,077	4412	834	11,129	81,985
Legnica	34,003	6697	1260	14,632	101,992
Wałbrzych	30,463	5223	960	12,311	117,926
Wrocław	244,584	68,170	20,399	105,257	632,067
OPOLSKIE	250,040	39,024	6553	74,806	1,004,416
Opole	53,308	13,159	2604	24,220	120,146
KUJAWSKO-POMORSKIE	541,972	96,040	21,466	155,136	2,092,564
Bydgoszcz	117,766	27,715	8340	44,882	359,428
Toruń	63,233	17,110	3816	24,406	203,447
Włocławek	31,747	5725	1208	10,556	114,885
POMORSKIE	571,154	128,917	29,566	182,060	2,295,811
Słupsk	26,527	5975	1653	11,102	93,936
Gdańsk	153,384	42,953	13,057	61,739	461,531
WARMIŃSKO-MAZURSKIE	334,732	52,757	9684	104,970	1,446,915
Elbląg	27,939	5355	1089	11,829	122,899
Olsztyn	63,414	15,542	3963	29,503	174,675

* group I includes wholesale and retail trade, servicing of private vehicles, transport and storage, hospitality and gastronomic services, information and communication services. ** group II includes insurance and financial services, real estate services. *** group III (other services) includes professional, scientific, and technical services, managerial and administrative services, public administration and national defense, obligatory social services, education, health and social care, cultural and recreational services, employment in households, production and services provided by households, foreign organizations. ****—for these cities (Sieradz, Ciechanów, Piła), no data were available for 2013; hence, these data are for 2012. Source: [29].

,

Table 4. The centrality measures of Polish cities in 2013 (corrected for population size).

Territorial Unit	The Centrality Measures of Polish Cities in 1998 (Corrected for Population Size)	Component Measures of Centrality
		I	II	III
ŁÓDZKIE
Łódź	96.8	29.9	11.1	55.8
Piotrków Trybunalski	62.9	19.6	1.6	41.8
Sieradz	49.3	5.1	0.1	44.1
Skierniewice	8.4	−10.1	−0.5	18.9
MAZOWIECKIE
Ciechanów	−21.8	−16.7	−12.5	7.4
Płock	13.7	1.6	−11.0	23.1
Ostrołęka	−19.6	−18.6	−12.8	11.8
Siedlce	1.4	−3.3	−16.4	21.1
Radom	−50.8	−23.4	−17.4	−10.0
Warsaw	203.6	69.7	39.0	95.0
MAŁOPOLSKIE
Cracow	167.1	58.3	17.3	91.5
Nowy Sącz	100.2	55.8	2.7	41.7
Tarnów	82.9	27.5	2.1	53.3
ŚLĄSKIE
Bielsko-Biała	87.1	38.2	0.5	48.4
Częstochowa	42.1	15.4	0.0	26.7
Katowice	267.2	83.0	42.1	142.1
LUBELSKIE
Biała Podlaska	63.4	33.8	2.5	27.1
Chełm	39.5	15.2	2.3	22.0
Zamość	101.3	40.0	7.6	53.7
Lublin	155.7	48.7	17.4	89.5
PODKARPACKIE
Krosno	178.3	78.6	5.7	94.0
Przemyśl	88.7	24.2	5.3	59.2
Rzeszów	204.3	71.8	17.6	114.9
Tarnobrzeg	41.0	2.0	4.1	34.9
PODLASKIE
Białystok	98.9	36.1	7.2	55.6
Łomża	60.2	19.6	3.2	37.3
Suwałki	41.6	17.9	7.2	16.5
ŚWIĘTOKRZYSKIE
Kielce	152.4	54.9	10.2	87.3
LUBUSKIE
Gorzów Wielkopolski	76.8	18.5	8.1	50.2
Zielona Góra	136.8	52.8	12.1	71.9
WIELKOPOLSKIE
Kalisz	43.2	17.3	0.6	25.3
Konin	58.3	19.0	2.8	36.5
Leszno	57.5	24.2	5.2	28.2
Piła	51.7	18.6	5.4	27.7
Poznań	182.3	57.7	20.4	104.2
ZACHODNIOPOMORSKIE
Koszalin	73.1	32.7	4.2	36.2
Szczecin	87.0	34.5	8.5	44.1
DOLNOŚLĄSKIE
Jelenia Góra	45.6	0.3	−1.2	46.5
Legnica	67.4	12.1	1.0	54.3
Wałbrzych	2.7	−9.2	−3.2	15.2
Wrocław	152.6	54.3	20.9	77.3
OPOLSKIE
Opole	212.9	70.7	15.1	127.1
KUJAWSKO-POMORSKIE
Bydgoszcz	94.9	31.2	12.9	50.7
Toruń	92.5	38.2	8.5	45.8
Włocławek	21.9	3.9	0.3	17.7
POMORSKIE
Słupsk	51.1	7.5	4.7	38.9
Gdańsk	106.8	36.9	15.4	54.5
WARMIŃSKO-MAZURSKIE
Elbląg	33.0	7.1	2.2	23.7
Olsztyn	164.9	52.5	16.0	96.4

Source: [29].

,

Table 5. The number employed in various service sectors in 2019 according to PBC 2007, together with population sizes.

Territorial Units	Total	Including:			Population
		I	II	III
ŁÓDZKIE	801,100	155,485	27,388	219,953	2,454,779
Łódź	256,231	67,934	18,838	105,627	679,941
Piotrków Trybunalski	24,562	4897	657	9911	73,090
Sieradz ****	12,290	2490	390	5411	42,762
Skierniewice	12,054	1958	448	4730	48,089
MAZOWIECKIE	2,030,021	523,378	144,493	680,108	5,423,168
Ciechanów ****	12,749	3339	493	5052	44,383
Płock	47,721	10,340	1418	17,784	119,425
Ostrołęka	17,919	3609	868	6804	52,055
Siedlce	27,433	6841	761	10,956	78,185
Radom	59,898	14,250	1587	23,675	211,371
Warsaw	1,002,743	315,803	124,898	424,853	1,790,658
MAŁOPOLSKIE	1,131,813	236,964	31,785	331,697	3,410,901
Cracow	376,552	114,590	21,930	169,449	779,115
Nowy Sącz	35,839	10,598	1131	10,911	83,794
Tarnów	41,840	9038	1042	16,255	108,470
ŚLĄSKIE	1,366,753	295,574	43,260	413,312	4,517,635
Bielsko-Biała	80,597	17,779	1934	26,343	170,663
Częstochowa	82,769	21,192	2033	26,317	220,433
Katowice	174,055	48,614	15,181	75,725	292,774
LUBELSKIE	709,509	94,377	14,641	172,017	2,108,270
Biała Podlaska	15,879	4273	489	6269	57,170
Chełm	15,112	3838	593	6313	61,932
Zamość	19,612	4630	803	8716	63,437
Lublin	133,258	34,643	7677	61,348	339,784
PODKARPACKIE	728,694	99,049	11,138	173,415	2,127,164
Krosno	23,889	4540	470	8120	46,291
Przemyśl	19,037	4680	607	8784	60,689
Rzeszów	99,181	24,701	4305	42,386	196,208
Tarnobrzeg	12,354	2225	437	5332	46,745
PODLASKIE	363,603	56,843	7560	96,017	1,178,353
Białystok	94,500	25,790	4011	42,217	297,554
Łomża	16,559	3637	487	7469	62,945
Suwałki	21,757	4536	724	7779	69,758
ŚWIĘTOKRZYSKIE	392,897	51,279	6134	101,686	1,233,961
Kielce	79,712	19,744	2716	36,805	194,852
LUBUSKIE	270,788	56,722	5917	79,063	1,011,592
Gorzów Wielkopolski	41,876	9222	1192	16,062	123,609
Zielona Góra	48,681	14,077	1891	19,939	141,222
WIELKOPOLSKIE	1,197,739	263,275	29,853	299,644	3,498,733
Kalisz	36,465	9342	858	11,881	100,246
Konin	24,856	6349	866	8901	73,522
Leszno	23,261	5341	962	7944	63,505
Piła ****	23,395	6623	827	7502	73,987
Poznań	257,595	71,938	15,019	117,396	534,813
ZACHODNIOPOMORSKIE	412,244	100,185	11,966	137,230	1,696,193
Koszalin	35,968	9588	1073	12,645	107,048
Szczecin	124,443	38,507	5671	54,629	401,907
DOLNOŚLĄSKIE	885,500	198,451	29,851	287,637	2,900,163
Jelenia Góra	23,514	4930	615	9560	79,061
Legnica	34,247	7530	1183	11,924	99,350
Wałbrzych	30,241	5774	745	10,784	111,356
Wrocław	295,011	88,324	18,799	132,603	642,869
OPOLSKIE	263,803	45,017	5317	76,505	982,626
Opole	61,054	14,728	2109	26,314	128,035
KUJAWSKO-POMORSKIE	584,111	116,536	19,923	164,357	2,072,373
Bydgoszcz	125,071	33,355	7042	47,524	348,190
Toruń	69,195	21,343	3645	27,394	201,447
Włocławek	31,995	6337	1085	9704	109,883
POMORSKIE	657,825	157,944	27,915	211,722	2,343,928
Słupsk	30,565	6238	1546	12,363	90,681
Gdańsk	183,773	57,925	11,992	76,379	470,907
WARMIŃSKO-MAZURSKIE	354,975	57,367	8171	110,046	1,422,737
Elbląg	28,513	5197	1062	11,869	119,317
Olsztyn	66,725	16,531	3055	32,646	171,979

****—for these cities (Sieradz, Ciechanów, Piła), no data were available for 2019; hence, these data are for 2016. Source: based on data from the Polish Central Statistical Office [41].

and

Table 6. The centrality measures of Polish cities in 2019 (corrected for population size).

Territorial Unit	The Centrality Measures of Polish Cities in 1998 (Corrected for Population Size)	Component Measures of Centrality
		I	II	III
ŁÓDZKIE
Łódź	118.9	36.6	16.5	65.7
Piotrków Trybunalski	47.5	3.7	−2.2	46.0
Sieradz	29.8	−5.1	−2.0	36.9
Skierniewice	−15.7	−22.6	−1.8	8.8
MAZOWIECKIE
Ciechanów	−48.4	−21.3	−15.5	−11.6
Płock	−1.2	−9.9	−14.8	23.5
Ostrołęka	−31.8	−27.2	−10.0	5.3
Siedlce	−11.2	−9.0	−16.9	14.7
Radom	−61.6	−29.1	−19.1	−13.4
Warsaw	234.8	79.9	43.1	111.9
MAŁOPOLSKIE
Cracow	216.7	77.6	18.8	120.2
Nowy Sącz	94.1	57.0	4.2	33.0
Tarnów	66.7	13.9	0.3	52.6
ŚLĄSKIE
Bielsko-Biała	103.4	38.7	1.8	62.9
Częstochowa	58.3	30.7	−0.4	27.9
Katowice	310.1	100.6	42.3	167.2
LUBELSKIE
Biała Podlaska	59.6	30.0	1.6	28.1
Chełm	40.2	17.2	2.6	20.3
Zamość	89.7	28.2	5.7	55.8
Lublin	171.8	57.2	15.6	99.0
PODKARPACKIE
Krosno	150.3	51.5	4.9	93.9
Przemyśl	98.5	30.6	4.8	63.2
Rzeszów	230.5	79.3	16.7	134.5
Tarnobrzeg	37.7	1.0	4.1	32.5
PODLASKIE
Białystok	105.9	38.4	7.1	60.4
Łomża	48.0	9.5	1.3	37.2
Suwałki	50.8	16.8	4.0	30.0
ŚWIĘTOKRZYSKIE
Kielce	175.2	59.8	9.0	106.5
LUBUSKIE
Gorzów Wielkopolski	74.1	18.5	3.8	51.8
Zielona Góra	114.2	43.6	7.5	63.0
WIELKOPOLSKIE
Kalisz	50.8	17.9	0.0	32.9
Konin	49.8	11.1	3.2	35.4
Leszno	54.9	8.9	6.6	39.4
Piła	32.7	14.3	2.6	15.8
Poznań	212.7	59.3	19.6	133.9
ZACHODNIOPOMORSKIE
Koszalin	70.7	30.5	3.0	37.2
Szczecin	98.8	36.7	7.1	55.0
DOLNOŚLĄSKIE
Jelenia Góra	13.2	−6.1	−2.5	21.7
Legnica	29.8	7.4	1.6	20.8
Wałbrzych	−22.5	−16.6	−3.6	−2.3
Wrocław	195.0	69.0	18.9	107.1
OPOLSKIE
Opole	207.9	69.2	11.1	127.7
KUJAWSKO-POMORSKIE
Bydgoszcz	107.4	39.6	10.6	57.2
Toruń	114.9	49.7	8.5	56.7
Włocławek	10.7	1.4	0.3	9.0
POMORSKIE
Słupsk	52.6	1.4	5.1	46.0
Gdańsk	141.0	55.6	13.6	71.9
WARMIŃSKO-MAZURSKIE
Elbląg	28.5	3.2	3.2	22.1
Olsztyn	180.3	55.8	12.0	112.5

Source: authors’ calculations based on data from the Polish Central Statistical Office [41].

. Note that current capitals are written in italics, while former capitals are indicated by normal script.

The numbers employed in various sectors in 1998 are presented in Table 1. The corresponding indices of centrality are presented in Table 2. The corresponding data for 2013 are presented in Table 3 and Table 4. Table 5 gives the number employed in various sectors in 2019, while Table 6 presents the corresponding corrected measures of centrality.

Table 7. Rankings of the global centrality measures in 1998, 2013, and 2019.

Rank 1998	City	${\mathit{W}}_{\mathit{C}}^{\mathit{M}}$	Rank2013	City	${\mathit{W}}_{\mathit{C}}^{\mathit{M}}$	Rank2019	City	${\mathit{W}}_{\mathit{C}}^{\mathit{M}}$
1	Katowice	179.6	1	Katowice	267.2	1	Katowice	310.1
2	Warszawa	170.4	2	Opole	212.9	2	Warszawa	234.8
3	Rzeszów	168.6	3	Rzeszów	204.3	3	Rzeszów	230.5
4	Krosno	160.5	4	Warszawa	203.6	4	Kraków	216.7
5	Opole	152.1	5	Poznań	182.3	5	Poznań	212.7
6	Poznań	136.4	6	Krosno	178.3	6	Opole	207.9
7	Olsztyn	132	7	Kraków	167.1	7	Wrocław	195.0
8	Nowy Sącz	124.6	8	Olsztyn	164.9	8	Olsztyn	180.3
9	Przemyśl	121	9	Lublin	155.7	9	Kielce	175.2
10	Kielce	119.3	10	Wrocław	152.6	10	Lublin	171.8
11	Sieradz	118.9	11	Kielce	152.4	11	Krosno	150.3
12	Zielona Góra	115.8	12	Zielona Góra	136.8	12	Gdańsk	141.0
13	Lublin	115.2	13	Gdańsk	106.8	13	Łódź	118.9
14	Kraków	113.8	14	Zamość	101.3	14	Toruń	114.9
15	Zamość	111.7	15	Nowy Sącz	100.2	15	Zielona Góra	114.2
16	Tarnów	103.2	16	Białystok	98.9	16	Bydgoszcz	107.4
17	Biała Podlaska	94.8	17	Łódź	96.8	17	Białystok	105.9
18	Leszno	92	18	Bydgoszcz	94.9	18	Bielsko-Biała	103.4
19	Białystok	90.9	19	Toruń	92.5	19	Szczecin	98.8
20	Piła	86.2	20	Przemyśl	88.7	20	Przemyśl	98.5
21	Wrocław	86	21	Bielsko-Biała	87.1	21	Nowy Sącz	94.1
22	Szczecin	80.6	22	Szczecin	87	22	Zamość	89.7
23	Toruń	78.6	23	Tarnów	82.9	23	Gorzów Wielkopolski	74.1
24	Gdańsk	77.2	24	Gorzów Wielkopolski	76.8	24	Koszalin	70.7
25	Chełm	76.9	25	Koszalin	73.1	25	Tarnów	66.7
26	Konin	74.7	26	Legnica	67.4	26	Biała Podlaska	59.6
27	Koszalin	74.3	27	Biała Podlaska	63.4	27	Częstochowa	58.3
28	Jelenia Góra	68.8	28	Piotrków Trybunalski	62.9	28	Leszno	54.9
29	Kalisz	67.3	29	Łomża	60.2	29	Słupsk	52.6
30	Gorzów Wielkopolski	67	30	Konin	58.3	30	Kalisz	50.8
31	Tarnobrzeg	66.9	31	Leszno	57.5	31	Suwałki	50.8
32	Piotrków Trybunalski	66.6	32	Piła	51.7	32	Konin	49.8
33	Skierniewice	66.5	33	Słupsk	51.1	33	Łomża	48.0
34	Bielsko-Biała	66.4	34	Sieradz	49.3	34	Piotrków Trybunalski	47.5
35	Bydgoszcz	64.9	35	Jelenia Góra	45.6	35	Chełm	40.2
36	Legnica	63.9	36	Kalisz	43.2	36	Tarnobrzeg	37.7
37	Suwałki	59.4	37	Częstochowa	42.1	37	Piła	32.7
38	Łódź	57.3	38	Suwałki	41.6	38	Legnica	29.8
39	Łomża	53.7	39	Tarnobrzeg	41	39	Sieradz	29.8
40	Słupsk	52.2	40	Chełm	39.5	40	Elbląg	28.5
41	Siedlce	48.8	41	Elbląg	33	41	Jelenia Góra	13.2
42	Elbląg	45.5	42	Włocławek	21.9	42	Włocławek	10.7
43	Częstochowa	41.3	43	Płock	13.7	43	Płock	−1.2
44	Włocławek	37	44	Skierniewice	8.4	44	Siedlce	−11.2
45	Ciechanów	34.8	45	Wałbrzych	2.7	45	Skierniewice	−15.7
46	Ostrołęka	25.5	46	Siedlce	1.4	46	Wałbrzych	−22.5
47	Wałbrzych	16.9	47	Ostrołęka	−19.6	47	Ostrołęka	−31.8
48	Płock	2.7	48	Ciechanów	−21.8	48	Ciechanów	−48.4
49	Radom	−20.1	49	Radom	−50.8	49	Radom	−61.6

Source: authors’ own calculations (for 2019); earlier figures are taken from [29].

presents the global measures of centrality in decreasing order. As in the previous tables, the current regional capitals are written in italics. It should be noted that in each case the majority of the top ten positions, including all of the top three positions, are occupied by current capitals. While three of the top ten positions were occupied by former regional capitals in 1998 (Krosno, Nowy Sącz, Przemyśl), in 2013, Krosno was the only former regional capital in the top ten. On the other hand, the lower positions are dominated by former regional capitals. This is particularly visible in 2019, when the bottom half of the table is occupied solely by former regional capitals.

Figure 2 illustrates the changes in the corrected values of the global index of centrality over the period from 1998 to 2019. It can be seen that the largest increases in this index correspond to current capitals. Only one current capital, Zielona Góra, exhibited a (small) decrease in the value of the global index of centrality. There exists a clear positive correlation between the size of a city and the global centrality index. The largest cities in Poland are Warsaw (1.8 mln. inhabitants), Cracow (0.8 mln.), Łódź (0.7 mln.), Wrocław (0.65 mln.), Poznań (0.54 mln.), and Gdańsk (0.5 mln.). The highest ranked city according to the increase in the global centrality index, Katowice (population 0.3 mln.), is not in itself amongst the largest cities in Poland. However, it is part of the Silesian conurbation whose total population is greater than that of the Warsaw conurbation. Only three of the former regional capitals exhibited an increase in this index over the period 1998–2019: Bielsko-Biała, Częstochowa, and Słupsk. Each of the three former capitals exhibiting the largest fall in the global index of centrality has a population of less than 50 thousand.

The Spearman coefficient of correlation between the population of a city in 1998 and the change in the global centrality index equals 0.8. Hence, it can be stated that there exists a strong, positive association between these two variables. While the importance of smaller cities in Poland as centers of employment has generally decreased, the relative importance of large cities in Poland has increased.

A more in-depth analysis of the current situation of Polish cities, together with the changes observed over the period 1998–2019, is presented in Section 4.1, Section 4.2, Section 4.3 and Section 4.4

4.1. Comparison of Mean Measures of Centrality According to Status

Due to the relatively small sample sizes (18 current capitals and 31 former capitals) and the skewness of the distribution of the measures of centrality, a non-parametric test (Wilcoxon test) was used to investigate systematic differences in the measures of centrality according to the status of a city. A significance level of 5% was applied; i.e., a difference was inferred to be statistically significant when the corresponding p-value was found to be less than 0.05.

(A)

In 1998

Note that, for example, the index 10⁻⁴ denotes that the first non-zero digit of the p-value appears in the fourth position after the decimal point (

Table 8. Comparison of the mean measures of centrality in 1998 according to the current status of a city. The standard deviations are given in brackets.

	Former Capital	Current Capital	p-Value
Global	67.84 (37.75)	111.43 (38.75)	4.152 × 10−4
Market Sectors (I + II)	34.31 (26.33)	76.96 (29.55)	1.152 × 10−5
Non-Market Sector (III)	33.54 (17.00)	34.47 (13.03)	0.6935

Source: authors’ own analysis.

).

Immediately prior to the administrative reforms, the current capitals exhibited a clearly higher level of centrality in the market sectors. The level of centrality in the non-market sectors was not associated with the status of a city resulting from the re-forms. Overall, the current capitals exhibited a higher level of centrality.

(B)

In 2013

In 2013, the current capitals exhibited a higher level of centrality in all of the sec-tors (

Table 9. Comparison of the mean measures of centrality in 2013 according to the current status of a city. The standard deviations are given in brackets.

	Former Capital	Current Capital	p-Value
Global	47.59 (43.24)	147.42 (53.22)	4.849 × 10−9
Sector I	14.71 (21.62)	49.98 (17.25)	3.743 × 10−6
Sector II	−0.12 (6.73)	16.66 (9.67)	9.192 × 10−9
Market Sectors (I + II)	14.59 (26.61)	66.64 (25.25)	9.034 × 10−9
Non-Market Sector (III)	33.01 (19.22)	80.77 (29.45)	4.662 × 10−7

Source: authors’ own calculations.

). The extremely low p-values suggest that the difference between the levels of centrality according to a city’s current status increased in the period from 1998 to 2013. This is confirmed by the following facts:

In this period, the measure of centrality for the market sectors clearly decreased in the former capitals, but decreased more slowly in the current capitals.

The measure of centrality for the non-market sectors did not significantly change in the former capitals, but clearly increased in the current capitals.

This will be analyzed more closely when the dynamics of change according to the current status of a city are investigated.

(C)

In 2019

In 2019, the current capitals exhibited a higher level of centrality in all of the sec-tors (

Table 10. Comparison of the mean measures of centrality in 2019 according to the current status of a city. The standard deviations are given in brackets.

	Former Capital	Current Capital	p-Value
Global	37.93 (46.43)	167.23 (62.05)	3.66 × 10−8
Sector I	8.97 (21.72)	57.02 (20.07)	1.46 × 10−7
Sector II	−0.90 (7.00)	15.66 (10.88)	2.016 × 10−8
Market Sectors (I + II)	8.07 (26.99)	72.68 (29.33)	3.611 × 10−10
Non-Market Sector (III)	29.86 (22.51)	94.57 (35.02)	1.039 × 10−7

Source: authors’ own calculations.

).

1. In this period, the measure of centrality for the market sectors again decreased in the former capitals, but increased almost back to 1998 levels in the cur-rent capitals.

2. The measure of centrality for the non-market sectors marginally decreased in the former capitals, but again increased in the current capitals.

In summary, at the beginning of the study period, the level of centrality in the non-market sectors was very similar among both the cities that lost their status as re-gional capitals and the current capitals. However, the level of centrality in the market sectors, and thus the overall level of centrality, was significantly higher in the current capitals. This indicates that, even before the administrative reforms, the current re-gional capitals had higher economic status than the cities that lost this status. Between 1998 and 2013, there was a fall in the level of centrality in the market sectors among both sets of cities, which was more prominent in the former capitals. While there was no significant change in the level of centrality in the non-market sectors among the former regional capitals, this level increased significantly amongst the current capitals. Between 2013 and 2019, the level of centrality in the market sectors amongst the cur-rent capitals returned to a similar level to that in 1998. On the other hand, this level continued to decline in the former capitals, albeit at a slightly slower rate. The level of centrality in the non-market sector continued to rise in the current capitals, while re-maining stable in the former capitals. Hence, although the pattern of changes is not uniform, the changes generally indicate an increasing disparity between the current regional capitals and the former capitals. These changes will be analyzed in more de-tail in the following subsections.

4.2. Dynamics of Change in the Former Capitals

The Wilcoxon test for paired values was used to test whether the mean centrality measures changed over the study period. Note that the mean values are all rounded to two decimal places, which may result in minor differences between the mean increase and the corresponding base values.

(A)

Between 1998 and 2013

Statistical analysis of the dynamics of change in the former capitals confirms the observations made above. In the period from 1998 to 2013, there was no significant change in the measure of centrality for the non-market sector (

Table 11. Changes in the mean measures of centrality for the 31 former regional capitals over the period from 1998 to 2013. The standard deviation of the changes is given in brackets.

	1998	2013	Mean Increase	No. of Cities with an Increase	p-Value
Global	67.84	47.59	−20.25 (22.18)	6	1.002 × 10−4
Market Sectors (I + II)	34.31	14.59	−19.72 (12.60)	2	1.304 × 10−8
Non-Market Sector (III)	33.54	33.01	−0.54 (16.36)	16	0.8848

Source: authors’ own calculations.

). On the other hand, there is a very clear fall in the measure of centrality for the market sectors, which corresponds to a clear fall in the global measure of centrality. The measure of centrality for the market sectors only grew in 2 of the 31 former capitals (Krosno and Płock). The changes in the non-market sector were evenly distributed: the centrality measure rose in 16 cities and fell in 15. The global change in centrality was positive in 6 of the 31 former capitals (Legnica, Łomża, Krosno, Częstochowa, Bielsko-Biała, and Płock).

(B)

Between 2013 and 2019

In the period from 2013 to 2019, the changes observed were very similar (

Table 12. Changes in the mean measures of centrality for the 31 former regional capitals over the period from 2013 to 2019. The standard deviation of the changes is given in brackets.

	2013	2019	Mean Increase	No. of Cities with an Increase	p-Value
Global	47.59	37.93	−9.66 (13.55)	7	5.411 × 10−4
Sector I	14.71	8.97	−5.74 (7.61)	6	2.477 × 10−4
Sector II	−0.12	−0.90	−0.78 (1.61)	9	0.01453
Market Sectors (I + II)	14.59	8.07	−6.52 (8.14)	5	1.672 × 10−4
Non-Market Sector (III)	33.006	29.858	−3.15 (10.52)	12	0.1671

Source: authors’ own calculations.

). The mean measure of centrality continued to fall in the market sector but did not change significantly in the non-market sector. The global measure increased in 7 of the 31 former capitals (Bielsko-Biała, Częstochowa, Chełm, Przemyśl, Suwałki, Kalisz, and Słupsk). It is notable that three of the cities (Chełm, Przemyśl, Suwałki) lie very close to the eastern border of Poland, a region that received a large amount of finance from EU structural funds. The measure of centrality for the market sectors increased in five of the cities (Nowy Sącz, Bielsko-Biała, Częstochowa, Chełm, Przemyśl). In each of these five cities plus Suwałki, the measure of centrality for Sector 1 increased in this period. The measure of centrality in Sector 2 increased in nine of the 31 former capitals (Os-trołęka, Nowy Sącz, Bielsko-Biała, Chełm, Konin, Leszno, Legnica, Słupsk, and Elbląg). The measure of centrality for the non-market sector increased in 12 of the 31 cities.

(C)

Over the Entire Study Period (1998 to 2019)

These results correspond to the results given above. The mean centrality measure for the former capitals in the non-market sector does not change significantly over the full study period. On the other hand, it falls significantly in the market sector, and as a result, the global mean measure of centrality also falls significantly. Over the entire period, the global measure of centrality grew in only 3 of these 31 cities (Bielsko-Biała, Częstochowa, and Słupsk). In the market sector, the measure of centrality only increased in Częstochowa. The measure of centrality in the non-market sector increased in 15 of the 31 former capitals (

Table 13. Changes in the mean measures of centrality for the 31 former regional capitals over the period from 1998 to 2019. The standard deviation of the changes is given in brackets.

	1998	2019	Mean Increase	No. of Cities with an Increase	p-Value
Global	67.835	37.932	−29.90 (27.91)	3	2.113 × 10−5
Market Sectors (I + II)	34.306	8.068	−26.24 (14.97)	1	1.825 × 10−6
Non-Market Sector (III)	33.542	29.858	−3.68 (19.48)	15	0.4102

Source: authors’ own calculations.

).

4.3. Dynamics of Change in the Current Capitals

(A)

Between 1998 and 2013

Both the global measure of centrality and the measure of centrality in the non-market sector increased in each of the 18 current capitals (

Table 14. Changes in the mean measures of centrality for the 18 current regional capitals over the period from 1998 to 2013. The standard deviation of the changes is given in brackets.

	1998	2013	Mean Increase	No. of Cities with an Increase	p-Value
Global	111.43	147.42	35.99 (21.43)	18	7.629 × 10−6
Market Sectors (I + II)	76.96	66.64	−10.32 (11.42)	4	0.002808
Non-Market Sector (III)	34.47	80.77	46.30 (19.90)	18	7.629 × 10−6

Source: authors’ own calculations.

). The measure of centrality in the market sectors increased in only 4 of these 18 cities (Łódź, Lublin, Wrocław, and Gdańsk).

(B)

Between 2013 and 2019

In this period, the global measure of centrality and the measure of centrality for the non-market sector continued to grow in a clear majority of the current capitals (

Table 15. Changes in the mean measures of centrality for the 18 current regional capitals over the period from 2013 to 2019. The standard deviation of the changes is given in brackets.

	2013	2019	Mean Increase	No. of Cities with an Increase	p-Value
Global	147.42	167.23	19.82 (18.27)	15	0.0006714
Sector I	49.98	57.02	7.04 (7.56)	15	0.002448
Sector II	16.66	15.66	−1.00 (2.68)	4	0.0883
Market Sectors (I + II)	66.64	72.68	6.04 (9.10)	14	0.01041
Non-Market Sector (III)	80.77	94.57	13.79 (10.77)	17	7.629 × 10−5

Source: authors’ own calculations.

). The previous fall in the measure of centrality in the market sector reversed. This is observed clearly in Sector 1. In the majority of current capitals, the measure of centrality in Sector 2 decreased. However, the change in this sector is not significant at a level of 5%. The measure of centrality in the non-market sector only fell in Zielona Góra (which shares the role of capital with Gorzów Wielkopolski). The global level of centrality decreased in Gorzów Wielkopolski, Zielona Góra, and Opole. These are three of the smallest regional capitals (the first two share this role in the Lubuskie region). The same pattern of change was also observed in Sector 1. The measure of centrality for Sector 2 increased in 4 of the 18 current capitals (Łódź, Warsaw, Cracow, and Ka-towice). These are among the largest cities in Poland.

(C)

Over the Entire Study Period (1998 to 2019)

Over the entire period, the measure of centrality in the non-market sector exhib-ited a very clear increase (in all of the current capitals), which in turn led to a clear in-crease in the mean global measure of centrality (

Table 16. Changes in the mean measures of centrality for the 18 current regional capitals over the period from 1998 to 2019. The standard deviation of the changes is given in brackets.

	1998	2019	Mean Increase	No. of Cities with an Increase	p-Value
Global	111.43	167.23	55.81 (34.82)	17	1.526 × 10−5
Market Sectors (I + II)	76.96	72.68	−4.28 (17.48)	8	0.3038
Non-Market Sector (III)	34.47	94.57	60.09 (26.61)	18	7.629 × 10−6

Source: authors’ own calculations.

). The only decrease in the global index of centrality was observed in Zielona Góra. The mean measure of central-ity in the market sectors did not change significantly overall (i.e., these measures re-turned to values similar to those observed in 1998 after an initial fall).

4.4. Comparison of the Dynamics of Change in Former and Current Capitals

(A)

Between 1998 and 2013

The increase in the measure of centrality for the non-market sector is clearly greater in the current capitals (where it grows significantly) than in the former capitals (where there is no significant change) (

Table 17. A comparison of the mean change in centrality measures according to the status of a city between 1998 and 2013. The standard deviation of changes is given in brackets.

	Mean Increase—Former Capitals	Mean Increase—Current Capitals	p-Value
Global	−20.25 (22.18)	35.99 (21.43)	3.256 × 10−8
Market Sectors (I + II)	−19.72 (12.60)	−10.32 (11.42)	0.01614
Non-Market Sector (III)	−0.54 (16.36)	46.30 (19.90)	2.568 × 10−8

Source: authors’ own calculations.

). The fall in the measure of centrality for the market sectors is significantly greater in the former capitals than in the current capitals. As a consequence, the increase in the global measure of centrality is significantly greater in the current capitals (where a clear rise is observed) than in the former capitals (where the mean measure falls).

(B)

Between 2013 and 2019

In the period from 2013 to 2019, the following mean measures of centrality exhibited greater increases in the current capitals than in the former capitals: (a) global, (b) for the market sectors, (c) for the non-market sector, (d) for Sector I. These differences were highly significant (p < 0.001 in all cases). On the other hand, there was no difference in the dynamics of change in Sector II according to the status of a city (

Table 18. A comparison of the mean change in centrality measures according to the status of a city between 2013 and 2019. The standard deviation of changes is given in brackets.

	Mean Increase—Former Capitals	Mean Increase—Current Capitals	p-Value
Global	−9.66 (13.55)	19.82 (18.27)	5.305 × 10−6
Sector I	−5.74 (7.61)	7.04 (7.56)	3.224 × 10−6
Sector II	−0.78 (1.61)	−1.00 (2.68)	0.3451
Market Sectors (I + II)	−6.52 (8.14)	6.04 (9.10)	3.507 × 10−5
Non-Market Sector (III)	−3.15 (10.52)	13.79 (10.77)	1.046 × 10−5

Source: authors’ own calculations.

).

(C)

Over the Entire Study Period (1998 to 2019)

The mean increase in the mean measure of centrality in the non-market sector over the whole study period was greater in the current capitals than in the former capitals (where there was no significant change in the mean measure of centrality) (

Table 19. A comparison of the mean change in centrality measures according to the status of a city between 1998 and 2019. The standard deviation of changes is given in brackets.

	Mean Increase—Former Capitals	Mean Increase—Current Capitals	p-Value
Global	−29.90 (27.91)	55.81 (34.82)	1.679 × 10−11
Market Sectors (I + II)	−26.24 (14.97)	−4.28 (17.48)	8.856 × 10−5
Non-Market Sector (III)	−3.68 (19.48)	60.09 (26.61)	5.193 × 10−12

Source: authors’ own calculations.

). The mean decrease in the mean measure of centrality in the market sector was greater in the former capitals than in the current capitals (where there was no significant change in the mean measure of centrality). As a result, the mean increase in the mean global measure of centrality was greater in the current capitals than in the former capitals (where there was a significant fall in the mean global measure of centrality).

These changes are clearly illustrated in Figure 3. In 1998, the overall centrality index in cities that retained their status as regional capital was significantly higher than in cities that lost this status. Since there were similar values of centrality in the non-market services according to status, this was determined by the clearly higher centrality index in the market services observed in cities that retained their status. In 2019, this difference was even larger, resulting from a dynamic increase in the centrality coefficient associated with non-market services in current capitals, and a decrease in the importance of market services in former capitals.

5. Discussion

In light of the results presented above, it can be stated that the evolutionary path of the central functions of Polish cities over the study period differed according to a city’s administrative role.

The group of cities whose central functions became less important (correcting for a city’s population) between 1998 and 2019 includes only those that lost their role as a regional capital. These former capitals can be generally described as large towns with a population below 100 thousand. The exceptions from this rule are Radom (210 thou.) and Wałbrzych (110 thou.), which have both faced major problems associated with the process of economic transformation and the collapse of local industry (coal mining in Wałbrzych and arms production in Radom) [42,43].

On the other hand, the group of cities whose central function became increasingly important predominantly includes current regional capitals. In addition, the largest cities were all ranked amongst those with the greatest increases in the importance of central functions. The largest increase was observed in Katowice (pop. 291 thou.)—which is the capital of the Silesia region and the largest city in the Silesian conurbation (pop. 3 mln.). Moreover, the cities of Bielsko-Biała (pop. 170 thou.) and Częstochowa (pop. 218 thou.), which are former regional capitals and now located in the Silesia region but at some distance from the Silesian conurbation, are two clear exceptions to this rule. Although these two cities lost their administrative status as regional capitals, the importance of their central functions has increased. Their demographic potential is sufficient to create their own development identity based on the industrial sector and the provision of services, both market (e.g., trade, tourism) and provided by the public sector (e.g., administration, health care, higher education) [2]. They provide an attractive alternative to the Silesian conurbation (Częstochowa is the largest city in Poland that does not have the status of a regional capital).

The cities that share the status of being the capital of a region—Zielona Góra (pop. 141 thou.) and Gorzów Wielkopolski (pop. 123 thou.) in the Lubuskie region, as well as Bydgoszcz (pop. 345 thou.) and Toruń (pop. 199 thou.) in the Kujawsko-Pomorskie region—have a less dynamic evolutionary path than cities of a similar size that independently play the role of a regional capital.

On the basis of such results, public debate has often drawn the conclusion that cities that lost their status as regional capitals are justified in seeing themselves as “losers” due to the administrative reforms in comparison to the cities that retained their status as regional capitals. This often leads to inferring the existence of a cause–effect relationship, according to which losing the status of regional capital automatically leads to the central functions of a city becoming less important. These intuitive conclusions have been confirmed by a number of academic studies. However, generally, such studies were conducted on the basis of opinion polls.

In the authors’ opinion, these results, in themselves, do not confirm the hypothesis that administrative reform influences the evolutionary path of a city’s development and that the loss of status directly inhibits such development. Closer inspection of the component elements in the general measure of centrality enables us to look at these changes from a wider perspective. In order to make a more accurate assessment of the nature of this phenomenon, the authors have analyzed the changes in the importance of central functions in the largest Polish cities. This analysis split the central functions of a city into public services and those provided by the market.

On the basis of this analysis, it was found that in 1998, immediately before the administrative reforms, the current regional capitals already exhibited a greater power of attraction than those cities that lost their status as regional capitals. This was mostly exhibited in a higher level of centrality in the field of services provided by the market. The importance of public services was at a similar level in both sets of cities. Hence, the greater attractive power of the current regional capitals is not simply the result of their administrative status. At the time of the administrative reforms, they already occupied a higher position in the hierarchy of cities. Hence, one may conclude that the administrative reforms of 1999 simply reflected the differing evolutionary paths of the cities that retained the status of regional capital and of those that lost this status. This is in line with the theory of the evolutionary paths of urban centers within a network of settlements [3,13,18].

Over the period from 1998 to 2019, the importance of public services (public administration and national defense, obligatory social services, education, health and social care) as an employer did not change significantly in the former regional capitals, while the importance of such services increased significantly in the current capitals. Hence, despite their loss of status, the importance of public services as an employer in former regional capitals did not decrease—as might be expected. On the other hand, the importance of the former regional capitals in providing employment within the market sector fell throughout the study period. On the other hand, while a fall in the importance of the market sector (e.g., trade, financial services) as an employer was initially observed in the current regional capitals, the importance of this sector has returned to a similar level to that in 1998. Hence, one might conclude that, in the field of providing market services, the former regional capitals no longer stand out from other towns and cities in the near vicinity. Hence, it is debatable whether the former regional capitals have become less attractive due to the administrative reforms. The current processes of socio-economic polarization, which concentrate economic growth in the largest urban settlements, can also be used to explain these evolutionary paths. This trend of metropolitanization is occurring, for the most part, independently of administrative reforms and not due to them. Hence, ascribing the degradation of Polish cities simply to administrative reform is unjustified.

It should be noted that the Polish administrative reforms were made in conjunction with the socio-economic transformation from a centrally planned economy to a neo-liberal one. Poland’s accession to the EU in 2004 accelerated the modernization processes and inflow of foreign capital. This was accompanied by a corresponding migration of highly educated young adults. This reinforced the polarizing trends in structural and spatial development. The service industries grew in importance, and this growth was concentrated in the largest cities. The gap between the largest cities and medium-sized cities increased, and as a result the former regional capitals (which tended to be smaller) naturally became more peripheral [44,45]. Researchers have also pointed out the importance of other factors that may have played a role in determining the path of development of Polish cities, such as an aging population [46], spatial differences in affluence [47], environmental problems [48], and the use of international connections by city administrations [49]. Our research indicates that the evolution of market services has played a crucial role in determining the evolution of a city’s global centrality measure. This supports the hypothesis that the process of metropolitanization has been more important in affecting the hierarchy of cities than administrative reforms. These results are in line with the conclusions made in earlier studies by Kurniewicz and Swianiewicz [50], Becker [33], and Heider, Rosenfeld, and Kauffmann [25]. These arguments concur with the theory of the evolutionary formation of an urban network, as described by a new economic geography model of central places [13]. This theory argues that economic integration promotes the growth of large cities, which attract an increasing number of firms and workers, at the expense of smaller cities.

6. Conclusions

The authors have attempted to address the research gap regarding the range of the effects of administrative reforms on the central functions of cities by analyzing economic data on employment. Due to the specific circumstances under which these administrative reforms were carried out, it is not possible to simply generalize these results to an international level. However, more broadly, the results of our research contribute many key insights to the global discussion of the effects of administrative reforms at the subnational level. The context in which reforms are made and a country’s level of economic development may have a significant influence on the effect of reforms on the evolution of the hierarchy of cities. This article has considered one specific context (Polish accession to the EU), and the question of how administrative reforms affect the evolution of the hierarchy of cities in other contexts is open to question. One possible direction for future research is to attempt to replicate these results in a variety of contexts.

Hence, this study has strengthened the evidence that the evolutionary path of the central functions of a city is complex and dependent on a large number of factors. Thus, the effect of administrative reforms on the hierarchy of cities is more difficult to predict than our intuition would suggest. Ascribing a causal role in the advance or degradation of a city to reform may well be an illusion, and the underlying reasons behind these development trends might lie in another set of processes. This aspect of the studied phenomenon might have very important practical implications—basing public administration policy on imperfect, although generally held, beliefs may well lead to suboptimal decision making. Hence, our study could be understood as part of the stream of research on data-driven Decision Support Systems (DSSs) [51].

The findings of this analysis provide valuable insights for policymakers, urban planners, and researchers striving to understand and manage the dynamic processes of urban transformation. The conducted study revealed that administrative reforms in Poland, particularly those implemented in 1999, significantly impacted the central functions of cities, both for those that lost their status as regional capitals and those that retained it. The analysis demonstrated that cities that lost their status experienced a decline in centrality, particularly in market sectors, which may suggest a deterioration in their role as local economic centers. However, some of these cities, through appropriate development strategies, managed to mitigate the negative effects and identify alternative growth paths. On the other hand, cities that retained their status showed an increase in centrality, highlighting the importance of maintaining administrative status for preserving and enhancing central functions. These results underscore the complexity of urban processes, indicating that administrative reforms can act as catalysts for change, but their effects are strongly dependent on local conditions and management strategies. Therefore, policymakers should consider these factors when planning future reforms and urban development to support sustainable growth and adaptation to changing socio-economic conditions.